dsf-cnn's Introduction

Dense Steerable Filter CNNs for Expoiting Rotational Symmetry in Histology Images

A densely connected rotation-equivariant CNN for histology image analysis.

Link to the pre-print.

NEWS: Our paper has now been published in IEEE Transactions on Medical Imaging. Find the published article here.

Getting Started

Environment instructions:

conda create --name dsf-cnn python=3.6
conda activate dsf-cnn
pip install -r requirements.txt

Repository Structure

src/ contains executable files used to run the model. Further information on running the code can be found in the corresponding directory.
loader/contains scripts for data loading and self implemented augmentation functions.
misc/contains util scripts.
model/class_pcam/ model architecture for dsf-cnn on PCam dataset
model/seg_nuc/ model architecture for dsf-cnn on Kumar dataset
model/seg_gland/ model architecture for dsf-cnn on CRAG dataset
model/utils/ contains util scripts for the models.
opt/ contains scripts that define the model hyperparameters and augmentation pipeline.
config.py is the configuration file. Paths need to be changed accordingly.
train.py and infer.py are the training and inference scripts respectively.
process.py is the post processing script for obtaining the final instances for segmentation.

Citation

If any part of this code is used, please give appropriate citation to our paper.

@article{graham2020dense,
  title={Dense Steerable Filter CNNs for Exploiting Rotational Symmetry in Histology Images},
  author={Graham, Simon and Epstein, David and Rajpoot, Nasir},
  journal={arXiv preprint arXiv:2004.03037},
  year={2020}
}